Reliability of high-speed railway bridges with respect to uncertain characteristics

The prediction of the dynamic response of railway bridges subjected to high-speed trains is investigated by stochastic methods to account for the unavoidable uncertainties. In this contribution Latin hypercube and line sampling is used for Monte Carlo simulations on a simple beam bridge model, to provide an overall overview of the structural reliability considering frequency variation, uncertain damping, and structural parameters. In particular, acceleration amplitudes define the limit states of the bridge models. Frequency uncertainties are a result of temperature induced stiffness changes and varying maintenance states of the bridge structure. With variation of the natural frequencies and bridge damping, both resonance speeds and the magnitudes of the resonance peaks are uncertain. Alternatively, on a more elaborated finite element bridge model, the influence of these uncertainties is investigated. The outcomes of the stochastic mechanical models with different degree of sophistication are compared with results of deterministic computations according to design guidelines.